Thousands of Exposed AI Servers Found Without Authentication
Trend Micro research discovered thousands of AI servers running popular open-source components exposed to the internet without authentication, including 200+ ChromaDB, 2,000 Redis, and 10,000+ Ollama servers. The exposures stem from zero-day vulnerabilities, rushed deployment practices, and insecure use of open-source and container-based infrastructure, creating potential risks of data theft, model poisoning, and ransom attacks.
What happened
Trend Micro's research found 200+ ChromaDB servers, 2,000 Redis servers, and 10,000+ Ollama servers exposed to the internet with no authentication.
The report identifies zero-day vulnerabilities and exploits in core AI components including ChromaDB, Redis, NVIDIA Triton, and NVIDIA Container Toolkit.
Reported impact
Trend Micro warns that exposed AI infrastructure creates risk of data theft, poisoning, ransom, and other attacks from threat actors.
Index analysis
The exposures resulted from rushed development and deployment timelines, as well as vulnerabilities in open-source components and container-based infrastructure used in AI systems.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.